Selective Small Molecule Activators of TREK-2 Channels Stimulate Dorsal Root Ganglion c-Fiber Nociceptor Two-Pore-Domain Potassium Channel Currents and Limit Calcium Influx

Abstract

The two-pore-domain potassium (K2P) channel TREK-2 serves to modulate plasma membrane potential in dorsal root ganglia c-fiber nociceptors, which tunes electrical excitability and nociception. Thus, TREK-2 channels are considered a potential therapeutic target for treating pain; however, there are currently no selective pharmacological tools for TREK-2 channels. Here we report the identification of the first TREK-2 selective activators using a high-throughput fluorescence-based thallium (Tl⁺) flux screen (HTS). An initial pilot screen with a bioactive lipid library identified 11-deoxy prostaglandin F2α as a potent activator of TREK-2 channels (EC₅₀ ≈ 0.294 μM), which was utilized to optimize the TREK-2 Tl⁺ flux assay (Z' = 0.752). A HTS was then performed with 76 575 structurally diverse small molecules. Many small molecules that selectively activate TREK-2 were discovered. As these molecules were able to activate single TREK-2 channels in excised membrane patches, they are likely direct TREK-2 activators. Furthermore, TREK-2 activators reduced primary dorsal root ganglion (DRG) c-fiber Ca²⁺ influx. Interestingly, some of the selective TREK-2 activators such as 11-deoxy prostaglandin F2α were found to inhibit the K2P channel TREK-1. Utilizing chimeric channels containing portions of TREK-1 and TREK-2, the region of the TREK channels that allows for either small molecule activation or inhibition was identified. This region lies within the second pore domain containing extracellular loop and is predicted to play an important role in modulating TREK channel activity. Moreover, the selective TREK-2 activators identified in this HTS provide important tools for assessing human TREK-2 channel function and investigating their therapeutic potential for treating chronic pain.

Keywords: DRG neuron; TREK-1; TREK-2; pain; thallium flux; two-pore-domain potassium channel.

Figure 1

Identification of TREK-2 bioactive lipid activators with a thallium flux assay. (A) Thallos fluorescence monitored before and after addition of Tl⁺ (arrow) to T-REx-TREK-2 cells induced with tetracycline (1 µg/mL, blue traces) or not induced (red traces). Inset is a Western blot run with T-REx-TREK-2 cell lysates with (+) or without (−) tetracycline induction and probed with a TREK-2 antibody. (B) Bioactive lipids that activate TREK-2 channels when preincubated with T-REx-TREK-2 cells for 5 min at a 10 µM concentration. (C) Tl⁺ flux (arrow) into thallos loaded TREK-2 expressing cells with (blue) or without (red) pretreatment with 11-deoxy prostaglandin F2α ± SEM (p < 0.001 after 300 s). (D) Dose–response curve for TREK-2 activation with 11-deoxy prostaglandin F2α ± S.E.M.; fitting determined an EC₅₀ value of 0.294 µM. (E) Calculation of the Z′ using the slope of Tl⁺ influx for each well of a 384 well plate containing TREK-2 expressing cells treated with 11-deoxy prostaglandin F2α (top) or vehicle (bottom, Z′ of 0.752).

Figure 2

Optimization of the secondary TREK-1 Tl⁺ flux assay. (A) Thallos fluorescence (F488) monitored before and after addition of Tl⁺ (arrow) to T-REx-TREK-1 cells induced with tetracycline (1 µg/mL, blue) or not induced (red) ± SEM. (B) Tl⁺ flux (arrow) into thallos loaded T-REx-TREK-1 expressing cells with (red) or without (blue) 10 min pretreatment with 1.8 µM 11-deoxy prostaglandin F2α ± SEM.

Figure 3

Top 10 TREK-2 activators. Concentration–response curves of compounds identified to activate TREK-2 channels measured as an increase in Tl⁺ flux and compared to TREK-1. Shown are the top 10 activators of TREK-2 with structures and estimated EC₅₀s ranked by potency (red traces) and compared to TREK-1 channels (blue traces). Each dose represented by a triplicate data point is expressed as normalized to a maximally effective dose of 11-deoxy prostaglandin F2α treatment (TREK-2) or Tl⁺ concentration (TREK-1). When displayed the EC50 was estimated using a nonlinear fit model with variable slope and normalized response in GraphPad Prism.

Figure 4

Top 10 TREK-2 inhibitors. Concentration–response curves of compounds identified to inhibit TREK-2 channels measured as decrease in Tl⁺ flux and compared to TREK-1. Shown are the top 10 compounds with structures ranked by potency for inhibition of TREK-2 channels (red traces) compared to TREK-1 channels (blue traces). Each dose represented by a triplicate data point is expressed as normalized to a maximally inhibitory dose of fluoxetine. When displayed, potency is an IC₅₀ estimated using a nonlinear fit model with variable slope and normalized response in GraphPad Prism.

Figure 5

TREK-2 activators are selective for TREK-2. (A) Representative TREK-2 channel activity during the indicated voltage ramp 5 min after addition of vehicle or 10 µM of the indicated TREK-2 activator. (B) Average change in TREK-2 current at 0 mV following 5 min after indicated treatment ±SEM (*p < 0.05). (C) The same voltage ramp experiments as in panels A and B were run on six other K+ channels (listed below the bar graph) before and after 10 µM of each TREK-2 activator; the resulting average change in current at 0 mV following each treatment is plotted ±SEM.

Figure 6

TREK-2 activators increase the channel’s open probability. (A) Inside-out voltage clamp recording of a TREK-2 channel at −30 mV with 11-deoxy prostaglandin F2α treatment (black bar). (B) Voltage ramp activation of TREK-2 channels with (blue) or without (red) 2 µM 11-deoxy prostaglandin F2α treatment. (C) Inside-out voltage clamp recording of a TREK-2 channel at the indicated voltages before and 5 min after the indicated treatment with T2A3. (D) Average open probability of TREK-2 at −30 mV before or 5 min after T2A3 treatment ± SEM (*p < 0.05).

Figure 7

A region on pore domain two is responsible for the activation of TREK-2 or inhibition of TREK-1 with 11-deoxy prostaglandin F2α. (A) Cartoon structures of the TREK-2 (gray)/TREK-1 (black) chimeric channels used to identify the region responsible for either activation or inhibition of TREK-2 or TREK-1, respectively. (B) Average change in Tl⁺ flux for the chimeric channels 5 min after addition of 2.5 µM 11-deoxy prostaglandin F2α ± SEM. (C) Tl⁺ flux change of the chimeric channels 5 min after addition of 10 µM T2A3 ± SEM. (D) Representative chimeric-1 TREK channel activity during the indicated voltage ramp 10 min after addition of vehicle (black trace) or 2.5 µM 11-deoxy prostaglandin F2α (red trace). (E) Representative chimeric-2 TREK channel activity during the indicated voltage ramp 10 min after addition of vehicle (black trace) or 2.5 µM 11-deoxy prostaglandin F2α (red trace). (F) Average change in chimeric channel current inhibition at −60, −30, and 60 mV following 10 min treatment with 2.5 µM 11-deoxy prostaglandin F2α ±SEM (n = 6 chimeric 1, n = 4 chimeric 2, *p < 0.05, **p < 0.01, ***p < 0.001). (G) Representative chimeric-3 TREK channel activity during the indicated voltage ramp 10 min after addition of vehicle (black trace) or 2.5 µM 11-deoxy prostaglandin F2α (red trace). (H) Representative chimeric-4 TREK channel activity during the indicated voltage ramp 10 min after addition of vehicle (black trace) or 2.5 µM 11-deoxy prostaglandin F2α (red trace). (I) Average change in chimeric current activation at −60, −30, and 60 mV following 10 min treatment with 2.5 µM 11-deoxy prostaglandin F2α ±SEM (n = 5 chimeric 3, n = 5 chimeric 4, n = 7 chimeric 5, *p < 0.05, **p < 0.01, ***p < 0.001).

Figure 8

Mouse DRG c-fiber TREK-2-like currents are activated by T2A3. (A) Mouse DRG currents recorded during a voltage ramp (−120 to 60 mV) before (red) and 5 min after vehicle (blue) treatment. (B) Mouse DRG currents recorded during a voltage ramp (−120 to 60 mV) before (red) and 5 min after T2A3 (blue) treatment. (C) Normalized increases in DRG K2P currents 5 min following vehicle or T2A3 treatments ± SEM (**p < 0.005).

Figure 9

TREK-2 activator T2A3 reduces DRG neuron Ca²⁺ influx. (A) Representative Ca²⁺ responses of IB4+ DRG neurons (inset of IB4+ staining) following 20 min pretreatment with DMSO (blue trace) or 10 µM T2A3 (red trace) and in response to the change in temperature (indicated above the traces). (C) The area under the Ca²⁺ curve for all DRG neuron Ca²⁺ responses at 25 and 40 °C ± SEM (n > 80 DRG neurons per group isolated from 4 mice, ***p < 0.001).